I offer services on projects related to web development, web design, and data science related projects such as computer vision and recommender systems.
Below are projects I have completed in these fields. If you have any specific projects you'd like to include, please let me know, and I can help incorporate them as well.
• Developed and optimized an end-to-end search-by-image system, integrating LangChain and ChromaDB with OpenAI’s CLIP embeddings to deliver accurate and scalable e-commerce product search and recommendations
• Spearheaded the design of advanced similarity computation pipelines, enabling real-time visual search and recommendation functionalities
Report• Developed a real-time car counting system using YOLOv8 for object detection and OpenCV for video processing, which ONLY detects and counts cars, separating them from other vehicles like vans and trucks
• Integrated the SORT algorithm to track and count only cars, ensuring precise vehicle monitoring across multiple video frames, even in dense traffic
• Optimized the processing pipeline for GPU execution with PyTorch, reducing latency by 50% and ensuring smooth performance for real-time video analysis
• Developed a real-time interactive balloon pop game using OpenCV and Pygame, enabling hand detection using CvZone for accurate interaction with in-game objects
• Achieved 90% hand detection accuracy over 500 frames, optimizing responsive interaction between the user's hand and in-game objects
• Developed a Content-Based job recommendation system using BERT embeddings to match user queries with relevant job postings based on semantic similarity
• Optimized the end-to-end data processing pipeline through PCA for dimensionality reduction and efficient cosine similarity calculations, ensuring scalability to handle over 100,000 job postings
• Implemented an HNSW algorithm for efficient nearest neighbor search, enabling quick retrieval of similar jobs from a large dataset
• Utilized Convolutional Neural Networks (CNNs) and computer vision techniques to design and implement a system for classifying tomato leaf diseases
• Achieved a high training accuracy of 97.65% and a low loss of 0.0718 which consistently delivered high confidence scores across multiple classes
• Integrated a CNN model into a React frontend and FastAPI backend web application for real-time disease detection, deployed on GCP to ensure reliability for multiple users, including farmers and agricultural professionals
Report• Developed a machine learning model for the accurate classification of diabetes using Naive Bayes and K-Nearest Neighbours algorithms to process and analyze clinical data
• Achieved a model accuracy of 75.5% with a maximum AUC score of 0.822, showcasing the model's robustness in clinical data analysis and prediction within healthcare AI applications
Report• Developed an autonomous GPT agent in Rust that writes, tests, and rewrites code, leveraging the OpenAI API and Actix Web
• Created special AI functions for extracting desired responses from large language models
• Developing a website for the NGO “Ruang Berubah,” aimed at connecting small and medium enterprises with buyers
• Currently progressing towards a final product resembling a detailed and secure Facebook Marketplace for small businesses
• Created a dynamic, responsive dashboard with real-time data visualization, updating every three months
• Designed and implemented front and backend of the data-leak register with a detailed form and dynamic table for managing data processing entries
• Managed routine database updates, conducted Database and Application testing and developed Symfony commands for automation
• Crafted a comprehensive online presence that highlights my professional accomplishments and personal journey since relocating abroad
• This website showcases my personal blog, representing the latest version of my web development journey
• To view my original personal portfolio website from when I first started learning web development, please click the button below
Old Website• Developed a desktop application designed for a product distribution platform, with the primary objective of enhancing company visibility and market reach efficiently
• Incorporated extensive backend database design using SQL to organize user information, order details, and product catalog
• Developed a real-time chat application using Firebase, showcasing skills in full-stack development and cloud computing. The app features user authentication, instant messaging, and responsive design.
• Engineered a state management using React Hooks and Zustand, ensuring efficient data handling and real-time updates for a seamless user experience.